Both theoretical and empirical labour market papers that incorporate t
he workers-hours dichotomy often contain assumptions about shapes of t
he overtime premium schedules faced by industry. Using cross-section d
ata for British production industries in three years of sharply contra
sting economic climates (1981, 1984 and 1988), this paper investigates
empirically the appropriate choice of schedule and, therefore, the in
dustrial cost consequences of changing the average number of overtime
hours per period. The analysis tackles problems associated with data d
isaggregation, the proportion of workers working overtime, omitted var
iables, heteroscedasticity and intertemporal variation.